Overview

  • Which normalizations of Hermite and Bessel polynomials are irreducible in every degree?
  • What do Stieltjes’ conjecture and the Legendre group conjecture assert?
  • How do opposite root pairs affect the full Galois group?
  • What is known about the nonzero roots of Hermite Wronskians?
  • Can distinct Legendre polynomials share a nonzero root?

Arithmetic properties

Orthogonal polynomials occur in approximation, differential equations, and spectral problems. Their recurrence and differential identities often give exact coefficients and simple real roots. Neither property alone forces irreducibility over Q\mathbb Q: a polynomial can have distinct real algebraic roots and still factor rationally.

Schur's work on Laguerre and Hermite polynomials established local arithmetic methods for such families. The all-degree irreducibility theorems for Hermite and Bessel polynomials are classical or published results. For Legendre polynomials, the general theorem is claimed in a recent preprint.

References 3

Hermite polynomials

For the physicists' Hermite convention Hn(x)=(−1)nex2(d/dx)ne−x2H_n(x)=(-1)^n e^{x^2}(d/dx)^n e^{-x^2}, one has

H2m(x)=(−1)m22mm!Lm(−1/2)(x2),H_{2m}(x)=(-1)^m2^{2m}m!L_m^{(-1/2)}(x^2),H2m+1(x)=(−1)m22m+1m!xLm(1/2)(x2).H_{2m+1}(x)=(-1)^m2^{2m+1}m!xL_m^{(1/2)}(x^2).

The odd polynomial has a forced factor xx; all remaining roots occur in opposite pairs. A theorem concerning Lm(±1/2)L_m^{(\pm1/2)} alone does not automatically determine the quadratic extensions obtained when square roots of its roots are adjoined.

Example

Here H4=16x4−48x2+12H_4=16x^4-48x^2+12. Its pair polynomial is 4y2−12y+34y^2-12y+3, with roots (3±6)/2(3\pm\sqrt6)/2. The pair field has degree two; the full splitting field also requires square roots of these roots. The same construction applies in higher degree.

Theorem (Schur)

For every m≥1m\ge1, the physicists' polynomials H2m(x)H_{2m}(x) and H2m+1(x)/xH_{2m+1}(x)/x are irreducible over Q\mathbb Q. This uses the normalization above. Rescaling the variable by an irrational number need not preserve irreducibility over Q\mathbb Q: for example, H2=4x2−2H_2=4x^2-2 is irreducible, whereas the probabilists' polynomial He⁡2=x2−1\operatorname{He}_2=x^2-1 is reducible.

References 3

Hermite Wronskians

Define the monic probabilists' polynomials by ext−t2/2=∑n≥0He⁡n(x)tn/n!e^{xt-t^2/2}=\sum_{n\ge0}\operatorname{He}_n(x)t^n/n!. For a partition λ=(λ1≥⋯≥λr>0)\lambda=(\lambda_1\ge\cdots\ge\lambda_r>0), put ni=λr−i+1+i−1n_i=\lambda_{r-i+1}+i-1 and

Wλ(x)=Wr⁡(He⁡n1,…,He⁡nr)∏i<j(nj−ni).W_\lambda(x)=\frac{\operatorname{Wr}(\operatorname{He}_{n_1},\ldots,\operatorname{He}_{n_r})}{\prod_{i<j}(n_j-n_i)}.

Write Wλ=xeλRλW_\lambda=x^{e_\lambda}R_\lambda, where eλe_\lambda is the multiplicity at zero. Grosu–Grosu prove that R(n,1)R_{(n,1)} is irreducible for n=1n=1 or n≥3n\ge3, R(n,2)R_{(n,2)} for n≥2n\ge2, and R(n,n)R_{(n,n)} when n+1n+1 is a square.

At n=2n=2, W(2,1)=x3W_{(2,1)}=x^3, so R(2,1)=1R_{(2,1)}=1, a unit, hence not irreducible. Their methods combine Newton-slope bounds with Schur-type congruences. Irreducibility implies simple nonzero roots, connecting arithmetic to differential equations. Irreducibility does not hold for every partition: R(6,3,2,1)=x2−9R_{(6,3,2,1)}=x^2-9. This concerns the displayed probabilists' normalization.

References 1
  • GrosuGrosu2021 · Definition 1, Theorems 1–3 and 7–8, and the discussion after Proposition 6 (arXiv:2007.00065v1)Source

Bessel polynomials

Theorem (Filaseta–Trifonov)

Let n≥1n\ge1, and let a0,…,an∈Za_0,\ldots,a_n\in\mathbb Z satisfy ∣a0∣=∣an∣=1|a_0|=|a_n|=1. Then

∑j=0naj(n+j)!2j(n−j)!j!xj\sum_{j=0}^n a_j\frac{(n+j)!}{2^j(n-j)!j!}x^j

is irreducible over Q\mathbb Q. Taking every aj=1a_j=1 proves the irreducibility of the Bessel polynomial yny_n in every positive degree.

The method bounds possible factor degrees by Newton polygons and prime divisors of factorial products. The valuation bounds allow the stated integral multipliers on the intermediate coefficients.

Example

The cubic Bessel polynomial is 15x3+15x2+6x+115x^3+15x^2+6x+1. Modulo 22 it is x3+x2+1x^3+x^2+1, which has no root and is therefore irreducible. Its discriminant is −135-135, a nonsquare, so its Galois group is S3S_3.

References 1

Generalized Bessel polynomials

Banerjee–Bera study irreducibility and Galois groups for a generalized Bessel family. Their 2020 paper has a 2021 corrigendum, which must be consulted with the original. These references concern additional parameters and do not change the all-degree theorem for the ordinary Bessel polynomials above.

References 2

Legendre polynomials

Define Pn(x)=12nn!(d/dx)n(x2−1)nP_n(x)=\frac1{2^n n!}(d/dx)^n(x^2-1)^n. Since Pn(−x)=(−1)nPn(x)P_n(-x)=(-1)^nP_n(x), the factor xx is forced at odd indices. Stieltjes' conjecture, as presented by Cullinan–Hajir, asks whether P2mP_{2m} and P2m+1/xP_{2m+1}/x are irreducible over Q\mathbb Q for every m≥1m\ge1. Their Galois-group questions concern these even polynomials and the corresponding signed-permutation action.

Example

After scaling, 8P4=35x4−30x2+38P_4=35x^4-30x^2+3, which is Eisenstein at 33. Its pair polynomial 35y2−30y+335y^2-30y+3 has discriminant 480480. Thus its pair field is Q(30)\mathbb Q(\sqrt{30}). This calculation proves irreducibility in degree four only.

Finite check

Exact modular certificates for the pair polynomials, together with norm witnesses for their quadratic lifts, verify that Pn(x)/xn mod 2P_n(x)/x^{n\bmod2} is irreducible over Q\mathbb Q for 2≤n≤2552\le n\le255. An independent verifier reconstructed the coefficients from the Legendre recurrence and checked the certificates. This finite range does not establish an all-degree theorem or determine the Galois groups.

References 3
  • CullinanHajir2014 · §1Stieltjes’ irreducibility conjecture and parity-reduced Galois groups.Source
  • CullinanHajir2014 · Lemma 1.3Norm criterion used to pass from pair irreducibility to lift irreducibility. The stated finite range was separately checked for this atlas on 12 November 2026.Source
  • Deng2026Legendre · Lemma 6.2, arXiv:2609.22336v2The same finite range, 2≤n≤255, is stated here. The atlas checked this range independently by modular and norm certificates; it did not verify the preprint’s all-degree argument or Galois groups.Source

Connections and evidence

Published

Legendre values: from prime divisibility to base-p digits

A published all-degree formula expresses vp(Pn(p))v_p(P_n(p)) through central binomial coefficients and gives a recurrence that reads the base-p digits of n.

Scope. Standard Legendre polynomials, every odd prime pp, and every integer n≥0n\ge0. The evaluation point is the same prime p as the valuation; the distinct p=2 formula is stated below.

Legendre values: from prime divisibility to base-p digitsA published all-degree formula expresses \(v_p(P_n(p))\) through central binomial coefficients and gives a recurrence that reads the base-p digits of n. Connection 1: Legendre values Pₙ(p), odd prime p is an input to A unique term of least valuation. The coefficient estimates identify one summand with strictly smaller p-adic valuation in the relevant parity expansion. Connection 2: A unique term of least valuation leads to Central-binomial valuation formula. The least-valuation term gives the displayed formulas for even and odd degrees without cancellation. Connection 3: Central-binomial valuation formula leads to Kummer: carries in m + m. The valuation of the central binomial coefficient equals the number of carries in the base-p addition m+m. Connection 4: Central-binomial valuation formula leads to Factorial valuations and digit sums. Writing factorial valuations as digit sums gives the all-degree digit formula in Theorem 4. Connection 5: Factorial valuations and digit sums leads to Exact base-p digit recurrence. Splitting n into its last base-p digit and remaining prefix yields the two-case recurrence. Connection 6: Kummer: carries in m + m is an input to Exact base-p digit recurrence. The carry description interprets the same valuation that the recurrence computes; it is not an extra hypothesis. Connection 7: Exact base-p digit recurrence leads to Compute the valuation digit by digit. The recurrence evaluates fₚ(n) using the digits of n and the parity of each prefix. Connection 8: Legendre values Pₙ(p), odd prime p leads to At p = 3: central Delannoy numbers. The binomial identity Pₙ(3)=Σⱼ binom(n,j)binom(n+j,j) identifies the central Delannoy sequence.Legendre values Pₙ(p), odd prime p — related articleLegendre values Pₙ(p),odd prime pA unique term of least valuation — related articleA unique term of leastvaluationCentral-binomial valuation formula — related articleCentral-binomialvaluation formulaKummer: carries in m + m — related articleKummer: carries in m + mFactorial valuations and digit sums — related articleFactorial valuations anddigit sumsExact base-p digit recurrence — related articleExact base-p digitrecurrenceCompute the valuation digit by digit — related articleCompute the valuationdigit by digitAt p = 3: central Delannoy numbers — related articleAt p = 3: central DelannoynumbersConnection 1: The coefficient estimates identify one summand with strictly smaller p-adic valuation in the relevant parity expansion.Expand the coefficients[1]Connection 2: The least-valuation term gives the displayed formulas for even and odd degrees without cancellation.Theorems 3–4 [2]Connection 3: The valuation of the central binomial coefficient equals the number of carries in the base-p addition m+m.Kummer’s carry formula[3]Connection 4: Writing factorial valuations as digit sums gives the all-degree digit formula in Theorem 4.Legendre’s factorialformula [4]Connection 5: Splitting n into its last base-p digit and remaining prefix yields the two-case recurrence.Theorem 6 [5]Connection 6: The carry description interprets the same valuation that the recurrence computes; it is not an extra hypothesis.Digitwise interpretation[6]Connection 7: The recurrence evaluates fₚ(n) using the digits of n and the parity of each prefix.Start from fₚ(0) = 0 [7]Connection 8: The binomial identity Pₙ(3)=Σⱼ binom(n,j)binom(n+j,j) identifies the central Delannoy sequence.Specialize p = 3 [8]
Linked nodes open articles; numbered connections open the evidence below.Open full-size diagram

deduction or information flow

Let fp(n)=vp(Pn(p))f_p(n)=v_p(P_n(p)), where p≥3p\ge3 is prime and Pn(1)=1P_n(1)=1. For every m≥0m\ge0, the published theorem gives

fp(2m)=vp(2mm),fp(2m+1)=1+vp(2m+1)+vp(2mm).f_p(2m)=v_p\binom{2m}{m},\qquad f_p(2m+1)=1+v_p(2m+1)+v_p\binom{2m}{m}.

Kummer identifies the binomial valuation with carries in m+mm+m in base pp. Equivalently, if sp(n)s_p(n) is the sum of the base-pp digits,

fp(n)=2sp(⌊n/2⌋)−sp(n)+(n mod 2)pp−1.f_p(n)=\frac{2s_p(\lfloor n/2\rfloor)-s_p(n)+(n\bmod2)p}{p-1}.

For every n≥0n\ge0 and 0≤a<p0\le a<p, this yields the complete recurrence

fp(pn+a)={fp(n)+(n mod 2),a even,fp(n)+1−(n mod 2),a odd.f_p(pn+a)=\begin{cases}f_p(n)+(n\bmod2),&a\text{ even},\\f_p(n)+1-(n\bmod2),&a\text{ odd}.\end{cases}

At p=3p=3, Pn(3)P_n(3) is the central Delannoy number. Since 13=(111)313=(111)_3, the recurrence gives f3(1)=1f_3(1)=1, f3(4)=1f_3(4)=1, and f3(13)=2f_3(13)=2; indeed P13(3)=1409933619=9⋅156659291P_{13}(3)=1409933619=9\cdot156659291, with the last factor prime to 33.

The prime 22 has its own formula: v2(Pn(2))=(n mod 2)−v2(n!)v_2(P_n(2))=(n\bmod2)-v_2(n!). These are valuation theorems in every degree, not irreducibility criteria.

Inspect all 8 connections and exact references
Evidence for each connection
ConnectionStatement and source
1Legendre values Pₙ(p), odd prime p → A unique term of least valuationConstruction / input

The coefficient estimates identify one summand with strictly smaller p-adic valuation in the relevant parity expansion.

  • V1 · Alekseyev et al. (2026)Theorems 3–4; Lemmas 10–11 and §5, author manuscript dated 4 July 2025, pp. 4–9Publication verified in Research in Number Theory 12 (2026), article 52; theorem numbers refer to the open author manuscript.Primary source
2A unique term of least valuation → Central-binomial valuation formulaDeduction

The least-valuation term gives the displayed formulas for even and odd degrees without cancellation.

  • V1 · Alekseyev et al. (2026)Theorems 3–4; Lemmas 10–11 and §5, author manuscript dated 4 July 2025, pp. 4–9Publication verified in Research in Number Theory 12 (2026), article 52; theorem numbers refer to the open author manuscript.Primary source
3Central-binomial valuation formula → Kummer: carries in m + mDeduction

The valuation of the central binomial coefficient equals the number of carries in the base-p addition m+m.

4Central-binomial valuation formula → Factorial valuations and digit sumsDeduction

Writing factorial valuations as digit sums gives the all-degree digit formula in Theorem 4.

5Factorial valuations and digit sums → Exact base-p digit recurrenceDeduction

Splitting n into its last base-p digit and remaining prefix yields the two-case recurrence.

6Kummer: carries in m + m → Exact base-p digit recurrenceConstruction / input

The carry description interprets the same valuation that the recurrence computes; it is not an extra hypothesis.

7Exact base-p digit recurrence → Compute the valuation digit by digitDeduction

The recurrence evaluates fₚ(n) using the digits of n and the parity of each prefix.

8Legendre values Pₙ(p), odd prime p → At p = 3: central Delannoy numbersDeduction

The binomial identity Pₙ(3)=Σⱼ binom(n,j)binom(n+j,j) identifies the central Delannoy sequence.

Sources and scope checked .

A claimed proof of Stieltjes’ conjecture

Preprint claim

Zikang Deng’s September 2026 preprint claims a proof of Stieltjes’ irreducibility conjecture in every degree. Version 2 uses local and real estimates for a resultant. This is recorded as a preprint claim; the full proof has not been independently checked.

References 1
  • Deng2026Legendre · Theorem 1.1, arXiv:2609.22336v2 (27 September 2026)A preprint claims a complete resolution; this atlas has not independently verified the full proof.Source

Common roots of Legendre polynomials

Preprint theorem · Kadets–Mangoubi

If α≠0\alpha\ne0 is a root of PℓP_\ell, then

#{n≥0:Pn(α)=0}≤2ℓ.\#\{n\ge0:P_n(\alpha)=0\}\le2\ell.

Their argument interpolates related polynomials 2-adically and bounds the zeros of the resulting analytic functions. The source conjectures that two distinct Legendre polynomials have no common nonzero root. This is a different question from irreducibility of each parity-normalized polynomial. The proposed all-degree irreducibility theorem would imply the common-root conjecture; Kadets–Mangoubi explicitly record that proposed theorem as a claim.

Indeed, a common nonzero root of two irreducible normalized members would force them to have the same degree. Only consecutive even and odd indices can have that degree; consecutive Legendre polynomials are coprime by their three-term recurrence.

References 2
  • KadetsMangoubi2026Legendre · Conjecture 1.1, Theorem 1.2, and the remark in §1.5; arXiv:2611.09312v1, 6 November 2026Statement and version checked; the complete preprint proof has not been independently audited.Source
  • Deng2026Legendre · Theorem 1.1, arXiv:2609.22336v2The stronger irreducibility statement remains listed separately as a preprint claim.Source

Methods and parameter ranges

The coefficient products of these families give explicit valuations and Newton polygons. Parity gives an action on opposite root pairs, and discriminant formulas can distinguish alternating from symmetric groups on the reduced roots. To transfer an argument between families, the change of variable must be defined over the ground field.

For Jacobi polynomials, fixing the parameters and letting the degree grow is different from fixing the degree and specializing a parameter. The former uses uniform arithmetic estimates; the latter can use covers and rational points.

References 2

Trace forms

Let f∈Q[x]f\in\mathbb Q[x] be monic and irreducible, with roots θ1,…,θn\theta_1,\ldots,\theta_n, and let E=Q[x]/(f)E=\mathbb Q[x]/(f). The trace form is the quadratic form z↦Tr⁡E/Q(z2)z\mapsto\operatorname{Tr}_{E/\mathbb Q}(z^2). Put

sk=∑i=1nθik,Δt=det⁡(si+j)0≤i,j<t.s_k=\sum_{i=1}^n\theta_i^k,\qquad \Delta_t=\det(s_{i+j})_{0\le i,j<t}.

Then Δn=disc⁡(f)\Delta_n=\operatorname{disc}(f). When all these minors are nonzero, their successive ratios diagonalize the trace form. Cullinan–Hajir–Young’s 2026 preprint develops continued-fraction formulas for Jacobi power-sum minors (Theorem 13) and applies them to Hasse–Witt invariants (§4). These invariants enter embedding problems for central double covers of alternating groups. They provide arithmetic information beyond the discriminant; they do not establish irreducibility for arbitrary Jacobi parameters.

Version note. The displayed formula in Theorem 6 of v1 needs a normalization correction: at t=0t=0 it gives n/(α+β+2n+1)n/(\alpha+\beta+2n+1), whereas the definition in §1.2 gives D0=Δ1/Δ0=nD_0=\Delta_1/\Delta_0=n, with Δ0=1\Delta_0=1. We therefore do not reproduce or use that diagonal formula here.

References 1
  • CullinanHajirYoung2026 · §1.2, Theorems 6 and 13, and §4; version 1Theorem 13 and §4 give the scope summarized here. The v1 Theorem 6 normalization discrepancy is recorded explicitly; its displayed diagonal formula is not used. Rational specializations must avoid degree loss and denominator zeros.Source

Legendre Galois groups

Write P2m+δ(x)=xδQm,δ(x2)P_{2m+\delta}(x)=x^\delta Q_{m,\delta}(x^2), with δ∈{0,1}\delta\in\{0,1\}. Cullinan–Hajir conjecture, under irreducibility of the degree-mm reduced polynomial, that its group is SmS_m and that the group of P2m+δ/xδP_{2m+\delta}/x^\delta is C2≀SmC_2\wr S_m. The two conclusions concern different permutation actions. Their Theorem 1.9(a) verifies the first conclusion through m=109m=10^9, still assuming irreducibility. A 2026 preprint claims the full group conjecture; it has not been independently verified here.

References 2

Connections and evidence

Claimed · preprint

Legendre polynomials: irreducibility and all signed permutations

Two preprints claim all-degree irreducibility and the full group C2≀SmC_2\wr S_m, combining resultant estimates with local cycles and the action on opposite root pairs.

Scope. Over Q\mathbb Q, for every m≥1m\ge1 and δ∈{0,1}\delta\in\{0,1\}, with the standard normalization Pn(1)=1P_n(1)=1 and L2m+δ(x)=P2m+δ(x)/xδL_{2m+\delta}(x)=P_{2m+\delta}(x)/x^\delta. Both all-degree inputs remain preprint claims.

Legendre polynomials: irreducibility and all signed permutationsTwo preprints claim all-degree irreducibility and the full group \(C_2\wr S_m\), combining resultant estimates with local cycles and the action on opposite root pairs. Connection 1: Parity-normalized Legendre polynomial is an input to Resultant bounds and finite degrees. Deng compares local and real bounds for the odd part of the factor resultant, then handles the remaining degrees. Preprint claim. Connection 2: Resultant bounds and finite degrees leads to Irreducible in every degree. The claimed contradiction and finite-range completion give irreducibility of both parity-normalized families. Preprint claim. Connection 3: Irreducible in every degree leads to Transitive action on squared roots. Writing L(x)=Q(x²), irreducibility of L implies irreducibility of Q and hence transitivity on its m roots. Connection 4: Parity-normalized Legendre polynomial is an input to Tame cycles and prime estimates. The quadratic Jacobi transformation makes coefficient valuations and selected tame inertia cycles explicit. Preprint claim. Connection 5: Transitive action on squared roots is an input to AND. Transitivity is one input to the group calculation. Connection 6: Tame cycles and prime estimates is an input to AND. The local cycles and prime estimates supply the additional group-theoretic input. Connection 7: AND leads to Full Sₘ on the root pairs. The preprint combines the published finite range with its large-degree primitivity argument and cycle criterion. Preprint claim. Connection 8: Tame cycles and prime estimates is an input to Odd proper flips; small cases. For m≥4 a suitable inertia power flips an odd proper set of pairs; the proof treats m≤3 separately. Preprint claim. Connection 9: Full Sₘ on the root pairs is an input to AND. The full action on pairs is required before conjugating a pair flip. Connection 10: Odd proper flips; small cases is an input to AND. The pair-flip input, or the explicit small-degree calculation, supplies the missing sign changes. Connection 11: AND leads to Full C₂ ≀ Sₘ on all 2m roots. The pair action and sign-kernel argument together yield the claimed full wreath product. Preprint claim.Parity-normalized Legendre polynomial — related articleParity-normalizedLegendre polynomialResultant bounds and finite degrees — related articleResultant bounds andfinite degreesIrreducible in every degree — related articleIrreducible in everydegreeTransitive action on squared roots — related articleTransitive action onsquared rootsTame cycles and prime estimates — related articleTame cycles and primeestimatesANDANDFull Sₘ on the root pairs — related articleFull Sₘ on the root pairsOdd proper flips; small cases — related articleOdd proper flips; smallcasesANDANDFull C₂ ≀ Sₘ on all 2m roots — related articleFull C₂ ≀ Sₘ on all 2mrootsConnection 1: Deng compares local and real bounds for the odd part of the factor resultant, then handles the remaining degrees.Assume a factorization[1]Connection 2: The claimed contradiction and finite-range completion give irreducibility of both parity-normalized families.Deng: all-degree claim[2]Connection 3: Writing L(x)=Q(x²), irreducibility of L implies irreducibility of Q and hence transitivity on its m roots.Pass to squared roots[3]Connection 4: The quadratic Jacobi transformation makes coefficient valuations and selected tame inertia cycles explicit.Local Jacobi description[4]Connection 5: Transitivity is one input to the group calculation.Transitivity [5]Connection 6: The local cycles and prime estimates supply the additional group-theoretic input.Primitivity and cycles[6]Connection 7: The preprint combines the published finite range with its large-degree primitivity argument and cycle criterion.All degrees; Thm 5.9 [7]Connection 8: For m≥4 a suitable inertia power flips an odd proper set of pairs; the proof treats m≤3 separately.Lemmas 5.10–5.11 [8]Connection 9: The full action on pairs is required before conjugating a pair flip.Permute every pair [9]Connection 10: The pair-flip input, or the explicit small-degree calculation, supplies the missing sign changes.Generate the sign kernel[10]Connection 11: The pair action and sign-kernel argument together yield the claimed full wreath product.Lemma 5.5; Theorem 5.9[11]
Linked nodes open articles; numbered connections open the evidence below.Open full-size diagram

deduction or information flow preprint claimAND joint hypotheses

Preprint claims in every degree. Put L2m+δ(x)=P2m+δ(x)/xδ=Qm,δ(x2)L_{2m+\delta}(x)=P_{2m+\delta}(x)/x^\delta=Q_{m,\delta}(x^2), where m≥1m\ge1 and δ∈{0,1}\delta\in\{0,1\}. Deng’s version 2 claims that L2m+δL_{2m+\delta} is irreducible. Nguyen-Dang’s version 1 claims the complete group determination

Gal⁡(Qm,δ/Q)=Sm,Gal⁡(L2m+δ/Q)=C2m⋊Sm.\operatorname{Gal}(Q_{m,\delta}/\mathbb Q)=S_m,\qquad \operatorname{Gal}(L_{2m+\delta}/\mathbb Q)=C_2^m\rtimes S_m.

The latter group acts on 2m2m roots: SmS_m permutes opposite pairs, and C2mC_2^m exchanges their members independently. Irreducibility alone does not determine either group. The two AND nodes show the extra inputs required.

The resultant comparison excludes factors in large degree; a finite completion is part of the irreducibility claim. On the group side, local cycles first determine the action on pairs. For m≥4m\ge4, an odd proper pair flip and its conjugates generate every sign change; the three smaller pair degrees receive explicit checks.

Why the resultant bounds conflict

For n=2m+δn=2m+\delta, write 2v2(n!)Pn(x)/xδ=H(x2)2^{v_2(n!)}P_n(x)/x^\delta=H(x^2). Deng’s Lemma 2.2 makes every rational factor of 2v2(n!)Pn(x)/xδ2^{v_2(n!)}P_n(x)/x^\delta even, so every proper factorization descends to H=ABH=AB. For a nontrivial factorization, take 1≤k=deg⁡A≤m/21\le k=\deg A\le m/2, and let RoddR_{\mathrm{odd}} be the positive odd part of ∣Res⁡(A,B)∣|\operatorname{Res}(A,B)|. For n≥64n\ge64, the Wronskian identity and orthogonality give the upper bound below; Legendre’s differential equation and Chebyshev–Hadamard estimates give the lower bound:

mlog⁡4−log⁡ ⁣4(m+1)2m<log⁡Roddk<(m+2n−1)log⁡3.m\log4-\log\!\frac{4(m+1)^2}{\sqrt m}<\frac{\log R_{\mathrm{odd}}}{k}<\bigl(m+\sqrt{2n-1}\bigr)\log3.

These bounds contradict each other for n≥256n\ge256; the finite completion covers 2≤n≤2552\le n\le255 (Deng, Theorems 4.1, 5.1 and §6).

The source statements and these dependencies were checked. The full preprint proofs have not been independently verified, and the all-degree conclusions retain the status Claimed.

Inspect all 11 connections and exact references
Evidence for each connection
ConnectionStatement and source
1Parity-normalized Legendre polynomial → Resultant bounds and finite degreesConstruction / input · preprint

Deng compares local and real bounds for the odd part of the factor resultant, then handles the remaining degrees.

  • L1 · Deng (2026)Theorem 1.1; Lemma 2.2; Theorems 3.1, 3.5, 4.1 and 5.1; §6, arXiv:2609.22336v2 (27 September 2026)All-degree irreducibility is a preprint claim. Statements and the indicated proof structure were inspected; the proof was not independently certified.Primary source
2Resultant bounds and finite degrees → Irreducible in every degreeDeduction · preprint

The claimed contradiction and finite-range completion give irreducibility of both parity-normalized families.

  • L1 · Deng (2026)Theorem 1.1; Lemma 2.2; Theorems 3.1, 3.5, 4.1 and 5.1; §6, arXiv:2609.22336v2 (27 September 2026)All-degree irreducibility is a preprint claim. Statements and the indicated proof structure were inspected; the proof was not independently certified.Primary source
3Irreducible in every degree → Transitive action on squared rootsDeduction

Writing L(x)=Q(x²), irreducibility of L implies irreducibility of Q and hence transitivity on its m roots.

4Parity-normalized Legendre polynomial → Tame cycles and prime estimatesConstruction / input · preprint

The quadratic Jacobi transformation makes coefficient valuations and selected tame inertia cycles explicit.

  • L2 · Cullinan–Hajir (2014)Lemmas 1.2–1.3; Conjecture 1.5; Theorem 1.9(a)The published finite group verification assumes irreducibility and concerns the degree-m action on pairs.Primary source
  • L3 · Nguyen-Dang (2026)Theorem 1.2; Proposition 5.7; Lemma 5.5; Theorem 5.9 and Lemmas 5.10–5.11, pp. 31–34, Hexagon 2610.00468v1 (22 October 2026)The all-degree group conclusion is recorded as a preprint claim; the root-pair action and the full 2m-root action are separate steps.Primary source
5Transitive action on squared roots → ANDConstruction / input

Transitivity is one input to the group calculation.

6Tame cycles and prime estimates → ANDConstruction / input

The local cycles and prime estimates supply the additional group-theoretic input.

7AND → Full Sₘ on the root pairsDeduction · preprint

The preprint combines the published finite range with its large-degree primitivity argument and cycle criterion.

  • L2 · Cullinan–Hajir (2014)Lemmas 1.2–1.3; Conjecture 1.5; Theorem 1.9(a)The published finite group verification assumes irreducibility and concerns the degree-m action on pairs.Primary source
  • L3 · Nguyen-Dang (2026)Theorem 1.2; Proposition 5.7; Lemma 5.5; Theorem 5.9 and Lemmas 5.10–5.11, pp. 31–34, Hexagon 2610.00468v1 (22 October 2026)The all-degree group conclusion is recorded as a preprint claim; the root-pair action and the full 2m-root action are separate steps.Primary source
  • L4 · Jones (2014)Corollary 1.3, published versionA primitive group containing a nontrivial single cycle with at least three fixed points contains the alternating group.Primary source
8Tame cycles and prime estimates → Odd proper flips; small casesConstruction / input · preprint

For m≥4 a suitable inertia power flips an odd proper set of pairs; the proof treats m≤3 separately.

  • L3 · Nguyen-Dang (2026)Theorem 1.2; Proposition 5.7; Lemma 5.5; Theorem 5.9 and Lemmas 5.10–5.11, pp. 31–34, Hexagon 2610.00468v1 (22 October 2026)The all-degree group conclusion is recorded as a preprint claim; the root-pair action and the full 2m-root action are separate steps.Primary source
9Full Sₘ on the root pairs → ANDConstruction / input

The full action on pairs is required before conjugating a pair flip.

10Odd proper flips; small cases → ANDConstruction / input

The pair-flip input, or the explicit small-degree calculation, supplies the missing sign changes.

11AND → Full C₂ ≀ Sₘ on all 2m rootsDeduction · preprint

The pair action and sign-kernel argument together yield the claimed full wreath product.

  • L3 · Nguyen-Dang (2026)Theorem 1.2; Proposition 5.7; Lemma 5.5; Theorem 5.9 and Lemmas 5.10–5.11, pp. 31–34, Hexagon 2610.00468v1 (22 October 2026)The all-degree group conclusion is recorded as a preprint claim; the root-pair action and the full 2m-root action are separate steps.Primary source

Sources and scope checked .

Jacobi and Gegenbauer parameters

Preprint claims

Nguyen-Dang’s Jacobi preprint also treats standard Pn(a,b)P_n^{(a,b)} with fixed distinct nonnegative integers a,ba,b. Theorem 1.3 claims irreducibility and group SnS_n for every sufficiently large nn, with an effective parameter-dependent threshold.

For integer r≥0r\ge0, m≥1m\ge1, and δ∈{0,1}\delta\in\{0,1\}, set Gm,r,δ=C2m+δr+1/2(x)/xδG_{m,r,\delta}=C_{2m+\delta}^{r+1/2}(x)/x^\delta, using the usual Gegenbauer polynomial. Theorem 1.4 claims irreducibility of this degree-2m2m polynomial when m≥max⁡{100r,1000}m\ge\max\{100r,1000\}, and when m≥max⁡{8r,M0}m\ge\max\{8r,M_0\} for an effective absolute M0M_0. These extend the paper’s Legendre and Kaneko–Zagier claims. The full proofs have not been independently verified here.

References 1

Questions

ClaimedPolynomial families

Are P2mP_{2m} and P2m+1/xP_{2m+1}/x always irreducible over Q\mathbb Q?

Scope. The standard Legendre polynomials and all integers m≥1.

This is Stieltjes’ conjecture in the cited 2014 paper. Deng’s September 2026 preprint claims a complete proof. The proof has not been independently verified here, so the result is recorded as a preprint claim.

References 2
Discuss this question →
SettledPolynomial families

Are the Bessel polynomials irreducible in every positive degree?

Scope. yn(x)=∑j=0n(n+j)!xj/(2j(n−j)!j!)y_n(x)=\sum_{j=0}^n (n+j)!x^j/(2^j(n-j)!j!) over Q, n≥1.

Yes. The Filaseta–Trifonov theorem proves this and permits arbitrary integral intermediate coefficient multipliers with endpoint multipliers ±1.

References 1
Discuss this question →
ClaimedPolynomial families

Are the reduced and full Legendre groups SmS_m and C2≀SmC_2\wr S_m?

Scope. m≥1m\ge1, δ∈{0,1}\delta\in\{0,1\}, assuming irreducibility of the degree-mm pair polynomial over Q\mathbb Q.

This is the two-part group conjecture of Cullinan–Hajir. It concerns the Galois groups, separately from Stieltjes’ irreducibility conjecture. A 2026 preprint claims a complete proof.

References 2
Discuss this question →
OpenPolynomial families

Are all nonzero zeros of every Hermite Wronskian simple?

Scope. Wronskians of Hermite polynomials with distinct nonnegative indices; zeros over C\mathbb C.

Veselov’s conjecture permits a multiple zero at the origin. It asks whether the remainder is squarefree; it does not assert irreducibility. Grosu–Grosu prove it for partitions (n,k,k−1,…,1)(n,k,k-1,\ldots,1), with n≥k≥1n\ge k\ge1. Felder–Veselov–Nekrasov still state the general problem as open in 2026.

References 3
Discuss this question →
OpenPolynomial families

Which two-part Hermite Wronskians have irreducible nonconstant remainders?

Scope. The probabilists’ normalization R(n,m)R_{(n,m)}, where n≥m≥1n\ge m\ge1 and deg⁡R(n,m)>0\deg R_{(n,m)}>0.

Grosu–Grosu ask to extend their irreducibility results to further two-part partitions. Their results include m=1,2m=1,2 and the diagonal when n+1n+1 is a square. The constant remainder at (2,1)(2,1) is excluded. A blanket statement for all partitions would be false.

References 1
Discuss this question →
OpenPolynomial families

Can distinct Legendre polynomials share a nonzero root?

Scope. Integers m,n≥1m,n\ge1, m≠nm\ne n; the Legendre polynomials Pn(x)=(2nn!)−1(d/dx)n(x2−1)nP_n(x)=(2^n n!)^{-1}(d/dx)^n(x^2-1)^n.

The conjecture is that PmP_m and PnP_n have no common root other than the forced root 00 at odd indices. Kadets–Mangoubi prove in their version-1 preprint that a fixed nonzero root of PℓP_\ell belongs to at most 2ℓ2\ell members of the sequence. Their Conjecture 1.1 asks for the stronger pairwise statement. The separate all-degree irreducibility claim would imply it; its full proof has not been independently verified here.

References 2
Discuss this question →

Related topics

References

  1. Affektlose Gleichungen in der Theorie der Laguerreschen und Hermiteschen Polynome

    Schur, Issai. Journal für die reine und angewandte Mathematik, 165, 52–58, 1931.

    Read the source
  2. The irreducibility of the Bessel polynomials

    Filaseta, Michael and Trifonov, Ognian. Journal für die reine und angewandte Mathematik, 550, 125–140, 2002.

    Read the source
  3. On the Galois groups of Legendre polynomials

    Cullinan, John and Hajir, Farshid. Indagationes Mathematicae, 25(3), 534–552, 2014.

    Read the source
  4. Algebraic properties of a family of Jacobi polynomials

    Cullinan, John and Hajir, Farshid and Sell, Elizabeth. Journal de théorie des nombres de Bordeaux, 21(1), 97–108, 2009.

    Read the source
  5. Irreducibility of generalized Hermite–Laguerre polynomials III

    Laishram, Shanta and Shorey, Tarlok N. Journal of Number Theory, 164, 303–322, 2016.

    Read the source
  6. The irreducibility of some Wronskian Hermite polynomials

    Grosu, Codruţ and Grosu, Corina. Indagationes Mathematicae, 32(2), 456–497, 2021.

    Read the source
  7. On a generalization of a conjecture of Grosswald

    Banerjee, Pradipto and Bera, Ranjan. Journal of Number Theory, 216, 216–241, 2020.

    Read the source
  8. Corrigendum to “On a generalization of a conjecture of Grosswald” [J. Number Theory 216 (2020) 216–241]

    Banerjee, Pradipto and Bera, Ranjan. Journal of Number Theory, 222, 426–428, 2021.

    Read the source
  9. A proof of the irreducibility conjecture for Legendre polynomials

    Deng, Zikang. arXiv preprint, 2026.

    Read the source
  10. Root sharing and 2-adic interpolation for Legendre polynomials

    Kadets, Borys and Mangoubi, Dan. arXiv preprint, 2026.

    Read the source
  11. Specializations of one-parameter families of polynomials

    Hajir, Farshid and Wong, Siman. Annales de l’Institut Fourier, 56(4), 1127–1163, 2006.

    Read the source
  12. Hasse-Witt invariants for trace forms of Jacobi polynomials

    Cullinan, John and Hajir, Farshid and Young, Elisabeth. arXiv preprint, 2026.

    Read the source
  13. Arithmetic of Jacobi polynomials

    Nguyen-Dang, Khai-Hoan. Hexagon preprint, 2026.

    Read the source
  14. Zeros of Wronskians of Hermite polynomials and Young diagrams

    Felder, Giovanni and Hemery, Adrian D. and Veselov, Alexander P. Physica D: Nonlinear Phenomena, 241(23-24), 2131–2137, 2012.

    Read the source
  15. Harmonic Locus and Calogero-Moser Spaces

    Felder, Giovanni and Veselov, Alexander P. and Nekrasov, Nikita. Communications in Mathematical Physics, 407(2), Article 22, 2026.

    Read the source
  16. Primitive permutation groups containing a cycle

    Jones, Gareth A. Bulletin of the Australian Mathematical Society, 89(1), 159–165, 2014.

    Read the source
  17. On the p-adic valuations of values of Legendre polynomials

    Alekseyev, Max A. and Amdeberhan, Tewodros and Shallit, Jeffrey and Vukusic, Ingrid. Research in Number Theory, 12, Article 52, 2026.

    Read the source